Detuning of T18 after high power test

Jiaru Shi18.04.2012

Outline

• RF measurement of CERN-Build T18– Comparison before and after high power test– Detuning Analysis

• Old results– (From Juwen Wang and Toshiyasu Higo)

• Related simulation• Geometry Change and analysis

RF measurementBefore high power test

after high power test

CERN built T18

Faya Wang @ SLAC

Faya Wang @ SLAC

After high power testBefore high power test

Increased standing waveRegular cells are fine

Analysis using the code for tuning from bead-pull data

• See detuning of input matching cell• The detuning of last two cells are calculated from the standing wave patter,

or the reflection from the output matching. Details next slides. (Note: while tuning, these two cells are tuned to correct the standing wave).

0 5 10 15 20-3

-2

-1

0

1

2

3

4

5x 10

6

number of cells

df a

fter h

igh

pow

er te

stin

g (H

z)

Standing wave patternReflection (both phase and amplitude)

• The imaginary part of this reflection can come from detuning of the last regular cell cell(N-1). Decrease of frequency

• Or the matching cell: Cell(N) Increase of frequency:

• Important note: the phase advance between the last two cells is ~100deg. Not 120 deg!

• Note: imag part of reflection comes from detuning, while real part of reflection comes from unmatched coupling

0.02

0.04

0.06

30

210

60

240

90

270

120

300

150

330

180 0

afterbefore

before last regular cell(N-1)

before Matching cell(N)

0.02

0.04

0.06

30

210

60

240

90

270

120

300

150

330

180 0

Amplitude Measurement of T18-SLAC #1 Before and After High Power Test

11421.7 MHz at 21.32°C, N2 Before high Power test

11421.87 MHz at 20.4°C, N2 After high power test

Juwen Wang @ SLAC

1350 hours

11421.7 MHz at 21.32°C, N2 ore high Power test

11421.87 MHz at 20.4°C, N2 After high power test

Phase Measurement of T18-SLAC #1 Before and After High Power Test

Juwen Wang @ SLAC

Amplitude Measurement of TD18-SLAC Before and After High Power Test

11424.5 MHz at 21.46°C, N2 Before high Power test

11424.56 MHz at 21.1°C, N2 After high power test

Juwen Wang @ SLAC

Phase Measurement of TD18-SLAC Before and After High Power Test

16.5°

11424.5 MHz at 21.46°C, N2 ore high Power test

11424.56 MHz at 21.1°C, N2 After high power test

Select bead pulling frequencies based on the same measurement condition for both before and after high power test

Juwen Wang @ SLAC

Amplitude Measurement of T18-SLAC #1 Before and After High Power Test

11424.1 MHz at 20.02°C, N2 Before high Power test

11424.15 MHz at 20.4°C, N2 After high power test

T24Juwen Wang @ SLAC

Phase Measurement of T24-SLAC Before and After 800 Hours High Power Test

Select bead pulling frequencies based on the measurement condition to get 2π/3 phase advance for both before and after high power test

6º

11424.1 MHz at 21.1°C, N2 After high Power test

11424.1 MHz at 21.2°C, N2 Before high power test

Juwen Wang @ SLAC

Similar Standing-Wave pattern for T(D)18

T18-CERN T18-SLAC

TD18-SLAC T24-SLAC

Summary of the detuningT18 SLAC N1 TD18 SLAC T24 SLAC T18 CERN N2

Measured at SLAC SLAC SLAC CERN

Output matching

Standing Wave(SWR)

1.06 1.2 1.05 1.1

(reflection) 0.03, -30dB 0.1,-20dB 0.025, -32dB 0.05, -26dB

Estimate df if one cell detuned

2MHz 7MHz 2.5MHz 3MHz

from standing wave pattern

+F@N cell or -F@N-1 cell

+F@N cell or -F@N-1 cell

+F@N cell +F@N cell or -F@N-1 cell

Regular cells

Total phase shift -16 deg +6 deg

~df +1 MHz -0.3 MHz

Note: T(D)18 structures have similar design where the phase advance between last two cells is ~100 deg. Making it hard to tell at Nth or at N-1th cell. It seems at the last cell.

RF measurement before/after baking• T24 12G N1

– Same configration– Temperature df

calculated

Phase S21

• Delta f ~10kHz

S21 magdf ~30kHz

11.8856 11.8857 11.8858 11.8859-36.84

-36.82

-36.8

-36.78

f / GHz

Com

bined

tran

smis

sion

S21

/ dB

-2.7775

-2.7644

11.9927

12.0344 12.0345 12.0346 12.0347

-40.6

-40.4

-40.2

-40

-39.8

-39.6

f / GHzCom

bined

tran

smis

sion

S21

/ dB

• T24 12G N2, leak fix at mode launcher– With and without wire, no direct comparison– Phase compare N/A because change of RF flange

adapter– df < 50kHz, Reflection increase

100k grid

• CERN PSI N2 after baking, df<100kHz– (same situation)

TD18 post-HPT @CERN• TD18 post-HPT (High Power Test) analysis

– 1. Chamfer

From: Markus Aicheler

TD18 Chamfer100 um -1MHz

Chamfer TD18

20.00 40.00 60.00 80.00 100.00r_chamfer [um]

-0.13

0.00

0.13

0.25

0.38

0.50

0.63

0.75

(re(

Mod

e(1)

)-11

425M

Hz)

/1M

Hz

Ansoft LLC f0XY Plot 2 ANSOFT

Curve Info

(re(Mode(1))-11425MHz)/1MHzSetup1 : LastAdaptive

Fill the 100um Chamfer +1MHz

T18 chamfer100 um -0.5MHz

10.00 30.00 50.00 70.00 90.00r_chamfer [um]

0.05

0.15

0.25

0.35

0.45

0.50

(re(

Mod

e(1)

)-11

424M

Hz)

/1M

Hz

Ansoft LLC cell_1XY Plot 1 ANSOFT

Curve Info

(re(Mode(1))-11424MHz)/1MHzSetup1 : LastAdaptive

Electromagnetic field

• Scaled to 150 MV/m Eacc• P = (-epsilon0 E^2 + mu0 H^2)/4• static simulation• Material: Copper E = 110GPa• Max deform: 0.06um, very

small. • 0.06um 12kHz• not the right direction

• HFSS result: Iris deform 10um ~ 2MHz

Circular Wg

Matching cell

E field pull B field

push

Asymmetry heating?

Circular Wg

Matching cell

Surface heating expansion?

• The source of detuning is not clear, but we find that. The MHz detuning corresponds to a geometry change that can be measured!

• The structure is cut open and critical dimensions are measured

T18 CERN N2 Cut

T18 CERN N2 Cut• 1: measure the profile of tuning bump,

compare with the tuning history

Tuning bump visible by eye

Height of tuning bump v.s. recorded tuning

• Good agreement “< +/- 1MHz”

11 12 13 14 15 16 17 18 19 200

1

2

3

4

5

6

7

8

0

10

20

30

40

50

60

70

80

90

delta f tuned (MHz)

average bump height (um)

Cell #

T18 CERN N2 Cut• 2: measure the distance between irises.

Distance between irises (gap)

12 13 14 15 16 17 18 19 20

-0.02

-0.015

-0.01

-0.005

0

0.005

g - g_RFdesign

Leftcenterright

mm

12 13 14 15 16 17 18 19 20

-0.025

-0.02

-0.015

-0.01

-0.005

0

0.005

g - g_RFdesign

leftcenterright

mm

Half-A Half-B

Error +-3 um. Three points are measured:“left” and “right” close to the cell wall, “center” close to the iris regionLast cell is longer in “center”.

12 13 14 15 16 17 18 19 20

-0.01

-0.005

0

0.005

0.01

0.015

0.02

0.025

0.03

g(center) - average(g(left,right))

Half-AHalf-B

mm

Simulation with deformed output matching iris in HFSS

• Single cell simulation gives 3MHz detuning from ~15 um “iris deform”.

• Full structure simulation shows similar standing wave pattern.

0.00 50.00 100.00 150.00 200.00Distance [mm]

0.00E+000

2.00E+004

4.00E+004

6.00E+004

8.00E+004

1.00E+005

1.20E+005

Com

plex

Mag

_E [V

_per

_met

er]

HFSSDesign1XY Plot 1 ANSOFT

Curve Info

ComplexMag_ESetup1 : LastAdaptivedz_iris='0.01mm' Freq='11.424GHz' Phase='0deg'

Summary (1)

• It’s a critical issue for the reliability and the lifetime of the accelerating structures.

• Almost every structure has an increased reflection from output, causing a standing wave pattern. For the same type of structure, the patterns are very similar.

• Geometry measurement on the T18 (CERN N2) shows deformation on the output matching iris This explains the increased reflection. (Most probably, this is also the case for the other 3 structures)

• In CLIC nominal design with compact coupler from the side, there will be no iris with such field asymmetry.

• Input side, cut into halves? Or take it iris by iris. Measure the profile of the whole input matching iris

• multiple-physics-coupled simulation: the deform of matching iris and the asymmetrical pulsed heating.

Summary (2)• Frequency change of regular cells is observed

in several structures, in TD18, but not in T18s; in T24s. in TD24?

• To be analyzed in near future: TD24 taken out from TBTS.– RF measurement, Cut, and dimensional control.